With the latest release of ChatGPT taking the world by storm, everyone is excited and concerned about this advancement in technology. In this article we will explore some concepts like Generative AI, what it can do for fashion and touch upon some of its limitations as well.
What is Generative AI?
Generative AI is a branch of artificial intelligence that uses training models and vast amounts of data to learn patters and characteristics and uses this to generate new and original content that resembles the training data.
It has been used to generate new content, such as text, images, music, or videos. What makes it powerful is that it can generate content keeping the context in mind. But it is important to keep in mind that Generative AI does not have inherent creativity as we humans do. Its output is based on the patterns that exist in the training data set. So, if the dataset is skewed you will get very biased results.
Generative AI is ushering in a new era of content generation in the fashion industry. From inspiring design innovation to providing personalized recommendations, enabling virtual try-on experiences, optimizing supply chains, promoting sustainability, and forecasting trends, generative AI is transforming every aspect of the fashion landscape.
ChatGPT is the current poster child for Generative AI.
What is ChatGPT?
This is how ChatGPT defines itself.
“ChatGPT is an advanced language model developed by OpenAI. It is based on the GPT (Generative Pre-trained Transformer) architecture and trained on a vast amount of text data from the internet and other sources to understand and generate human-like text responses. It can engage in conversations, answer questions, provide explanations, and assist with various language-related tasks.
The model is designed to understand context, generate coherent responses, and mimic human conversation. It has the ability to generate text that is often indistinguishable from text written by a human. ChatGPT can be used for a wide range of applications, such as AI chatbots, virtual assistants, content generation, language translation, and more.
It's important to note that as an AI language model, ChatGPT doesn't have personal experiences or consciousness. It operates based on patterns and examples in the data it was trained on. It's a powerful tool for language processing but should be used with care and human oversight to ensure responsible and ethical use.”
The machine learning part is provided by the ‘GPT’ in ChatGPT, which stands for ‘Generative Pre-Trained Transformer.’ This can be thought of as a digital emulation of a human brain
Its language skills are fantastic. It performs a large variety of language processing tasks extraordinarily well. Using simple conversational language, it responds with contextual understanding without grammatical errors.
Why is the world of Fashion so excited by chatGPT?
With ChatGPT the fashion industry is poised to reach new heights of creativity, engagement, and success. Although AI and Machine learning technology has been applied in various ways in the fashion industry, the quality of the conversational AI chatbot format makes it compelling. It is the perfect tool for fashion retailers obsessed with customer engagement.
The immediate applications of ChatGPT in fashion retail would be:
Fashion Advisory: ChatGPT already has access to crowd-sourced knowledge about Fashion and styling. It is shocking how good its responses are to general fashion queries
How to style a biker jacket?
I'm attending a summer wedding, what should I wear that is both stylish and breathable?
I have a pear-shaped body, what clothing styles would flatter my figure?
I do not like to show off my legs, suggest skirts that would work for me
Would a brown top look good on me if I have a dusky skin tone?
Stylebots - Shoppers can chat with the AI bot just like they would with a stylist or a sales assistant. Instead of browsing and scrolling through 100’s of products they can be taken on a guided shopping journey seeing only what is relevant for them. You can even handle customer support queries in a more supportive manner.
* Product recommendations - The inherent intelligence in ChatGPt also tells you what products will look good together or what to wear for a specific occasion. The challenge for retailers will be to retrieve these products from their inventory in real-time.
* Personalisation- While personal preferences, tastes, lifestyles etc are important there are other factors like body shape, skint one, weather, location, occasion etc which chatGPt is able handle effortlessly.
* Product descriptions - Writing creative descriptions for 100’s of product is really challenging. Now, Retailers can have diverse product descriptions for all their products almost instantaneously
* Language Translations - I have not tested this personally but apparently you can use ChatGPT to translate text into multiple languages, including English, Spanish, French, German and Chinese.
This is just scratching the surface. As our understanding of Generative AI gets better, I am sure more use cases will come up and we will use this technology in surprising ways.
But now for a few words of caution:
Limitation of ChatGPT
Bias: As mentioned earlier ChatGPt can have biases based on tits training data. It is only as good as its training data.
Reliability of source: ChatGPT is trained on crowd-sourced data. There is all kinds of data online, some incorrect and illogical as well. But ChatGPT will answer your question with authority although the information has not been verified.
Training data is not up to date- This is just a current limitation and am sure it will be fixed soon. But at this point ChatGPT has been trained on a dataset that only extends to 2021
No multi-media- Again I am sure this will change, and we will see images, audio and video outputs as well in the coming releases.
Hallucination:(This word is being in the technical sense, and I loved it) Since ChatGPt summarises all the info into original output, over time it will use its output as a source as well and can create an echo chamber of its own thoughts.
How we are using ChatGPT
Streamoid had launched a stylebot way back in 2017. But there were some challenges for large-scale applications as the interface was not conversational and had a lot of radio buttons etc.
With the latest version of ChatGPT we are able to take the best of its general intelligence and its conversational interface and add it to Streamoid Fashion intelligence.
Magix, the Stylebot from Streamoid is for a B2B applications where it will act as the brands stylists and make recommendations keeping the brands styling guidelines in mind. It also doubles up as a sales associate and help shoppers discover products and make smart purchases.
Streamoid’s Stylebot begins where the limitations of ChatGPT begin. It uses data from verified knowledge banks, can use rich media, and can offer truly personalized styling advice keeping individual preferences and tastes in mind.